Service for visual assessment of food preparation quality (ML/CV)

Budget: 100 EUR Deadline: 2 days

Hello,

I am ready to offer technical consultation on the development of a service for visual assessment of food preparation quality.

My approach:
• Use of computer vision (CV) and machine learning (ML) methods for analyzing images of dishes.
• Application of convolutional neural networks (CNN) for classifying quality levels (e.g., well-cooked / undercooked).
• Ability to supplement the model with visual aesthetics metrics (color, texture, evenness of cooking).

Technology stack:
• Python (TensorFlow / PyTorch)
• OpenCV for image preprocessing
• FastAPI or Flask for creating the service and REST API
• Ability to integrate with cloud platforms (Google Cloud, AWS) for scalability

Approximate algorithm:
1. Collection and annotation of a dataset of images of dishes with different quality levels.
2. Preprocessing of images (normalization, augmentation).
3. Training of the CNN model for quality classification.
4. Validation and testing of the model on new data.
5. Development of a service with an API for uploading images and receiving assessments.
6. Possibility of expanding functionality (e.g., recommendations for improving cooking).

Thanks to my international experience in AI/UX and working with various markets, I can offer a solution that will be not only technically correct but also user-friendly.

I am ready to discuss the details and adapt the solution to your requirements.

Sincerely,
Assontsa Sonfack Toscani Expand

Budget: 80 EUR Deadline: 3 days

Hello, I did a similar project that calculated the calories of a dish.

I would choose the simplest algorithm possible:
The service accepts images of the dish and sends it to gemini/chatgpt (I personally recommend gemini).

In the prompt, we describe how exactly we want to evaluate the dish. We describe the color/presentation/...

I assure you that the accuracy will be better than any other chosen algorithm. Training your own model is expensive. And the current versions of AI allow classifying anything, but a good prompt is additionally needed. This is the most effective algorithm.

Budget: 50 EUR Deadline: 1 day

I am among the top 5 developers in the category of "Artificial Intelligence and Machine Learning" among ~2100 specialists on the platform.
I guarantee:
- Fast and high-quality task execution
- Strict adherence to deadlines
- Regular communication throughout the entire process
I would be happy to discuss the details of your project in private messages.

Budget: 50 EUR Deadline: 3 days

Hello, I have been working with Python for over 5 years, and I have been working with OpenCV for several years. In this case, we need to develop a classifier. We can look for similar models and fine-tune them or create our own from scratch. There are ready-made models for dish classifiers, but to determine how a dish is prepared, we need to analyze colors and ingredients. Therefore, we need to create several stages and models: first, a dish classifier trained on a sufficiently large dataset; then we analyze the color palette of the dish and compare it with dishes from the dataset; the greater the difference, the worse the dish; next, we analyze the ingredients, checking if there are any recognizable products, their quantity and ratio, and compare them with the ideal values from the dataset. Overall, the task is interesting, and I would be happy to discuss the details!

Budget: 50 EUR Deadline: 1 day

Hello! I can consult you and turn your idea into a clear task statement, which can then be realistically assembled into a working MVP.

I have relevant experience with computer vision in complex production cases (I can share privately if you are interested). With tasks like "assessing quality from a photo," the most important thing is to correctly define what we consider quality, what features should be visible in the image, and how to organize the data so that the model does not learn from the background/light instead of the product.

I propose a dialogue in a format convenient for you, where we can quickly go through your dishes/categories, the source of the photo (one camera/different phones), the format of the response (accept/reject, classes or score), and the acceptable error. After that, I will send a brief document with the recommended approach (which architecture/stack, what data and labeling are needed, which metrics make sense, what risks there are, and what to do to eliminate them) and provide a realistic assessment of what accuracy can be achieved under your conditions and what volume of data will be required for this.

If after the consultation you want to move forward, I can also propose a plan for implementing the MVP (API/demo, basic model, quality check on a control test).

Budget: 130 EUR Deadline: 1 day

Hello,

Thank you for the detailed description, I would be happy to help with the implementation of the project. I am a Software Development Engineer with over 7 years of experience in creating various websites, SaaS solutions, complex web platforms, and MVPs for startups. In my portfolio, I have highlighted some of the projects I have developed.

My stack includes the following technologies and their frameworks: PHP (Yii2, Laravel, Symfony), HTML5, CSS3, JS (Vue.js, React.js), MySQL, PostgreSQL.

Budget: 1500 EUR Deadline: 10 days

Hi,

Methods
- Computer vision with deep learning
- CNN-based image classification/regression
- Transfer learning (pretrained food/image models)
- Optional: object detection/segmentation for complex dishes

Technology stack
- Python
- PyTorch or TensorFlow
- OpenCV (image preprocessing)
- FastAPI/Flask (backend API)
- Docker, GPU (CUDA)
- Cloud: AWS / GCP / Azure
- Optional frontend: Web or Mobile app

Approximate algorithm
- Collect and label dish images (undercooked / normal / overcooked or quality score).
- Preprocess images (resize, normalize, augment).
- Train a CNN model using transfer learning.
- Validate and tune the model.
- Deploy the model as an API service.
- User uploads a photo → model analyzes → returns cooking quality and confidence.

Best Expand